Resource & Links  

Review: Acupuncture for Stroke

Francine Rainone, PhD, DO, MS, Director of Community Palliative Care, Department of Family Medicine, Montefiore Medical Center, Bronx, NY

Introduction
It is now common for people to publish reviews of the literature concerning Complementary and Alternative Medicine (CAM) for specific conditions. In a typical review, exhaustive searches are completed and then criteria for scientific validity are applied, resulting in the exclusion of most of the studies. The remaining studies are subjected to statistical analysis, and the conclusion is often that there is
insufficient evidence to recommend for or against the therapy. While
such reviews are valuable in determining whether the efficacy of a
treatment has been established, their narrow focus rarely provides help
to guide clinicians already using the therapy or researchers interested
in furthering CAM research. My intention here is to utilize most of the
literature to review the topic of acupuncture for stroke. The purpose
of
the review is not to assess the evidence for efficacy so much as to
think through the issue as an acupuncturist and a researcher.

In the People's Republic of China, acupuncture is widely used as an
adjunctive therapy in the treatment of stroke. In the United States it
is almost never used. In general, novel therapies are investigated in
this country because they elicit physiological mechanisms that suggest
potential efficacy and/or because there is compelling clinical evidence
of their effectiveness. Numerous studies demonstrate that acupuncture
induces physiologic changes that promote healing after ischemic and
hemorrhagic brain injury, but the clinical evidence is mixed. A recent
review concluded that there is no additional benefit to administering
acupuncture as part of a comprehensive rehabilitation program following
stroke.(1) This conclusion is premature.

Several types of "clinical evidence: should be distinguished. The
lowest
level is anecdote. Anecdotal evidence, as defined here, refers to case
reports involving ten or fewer patients whose clinical improvement is
associated with a particular intervention. Anecdotal evidence is often
the start of off-label uses of pharmaceuticals or botanicals. Anecdote
can be an important source of innovation, but necessarily runs the risk
of leading to unforeseen consequences.  The next level of clinical
evidence is custom. The customary way of doing things is accepted as
useful regardless of the evidence base. Separating the customary from
the efficacious is difficult for every health care practitioner.

Physiologic Studies in Animals
At least seven animal studies of the effect of acupuncture following
transient brain ischemia have shown that acupuncture regulates chemical
mediators of ischemia. Much of the work has centered on nitrous oxide
(NO). Levels of NO, which is responsible for excitotoxicity, are
significantly increased during cerebral ischemia and reperfusion.
Acupuncture has been shown to regulate NO levels by several mechanisms.
In three studies in rats, 60 minutes of electroacupuncture (EA)
delivered to two points 10 minutes after occlusion of the middle
cerebral artery (MCA) inhibited overexpression of both neuronal and
inducible nitric oxide synthase messenger RNA (nNOS and iNOS
mRNA).(2-4)
A subsequent study showed that the same EA regimen inhibited NO release
by 53% at 56 minutes following MCA occlusion, though it did not
normalize levels.(5) In rats with chemically induced epilepsy, EA also
inhibited nNOS and iNOS to a statistically significant degree.(6)  In
addition to inhibiting the genetic precursors to _expression and the
release of NO, acupuncture affects its production by another mechanism.
Glial cell line derived neuronal factor (GDNF) reduces the volume of
infarct in rat brains by inhibiting neuronal NO production and neuronal
apoptosis. EA administered immediately after MCA occlusion did not
increase peak GDNF mRNA _expression in rat brains, but did delay the
decline of GDNF that occurs 12 hours after reperfusion.(7)

Studies also document the effects of acupuncture on regulators and
mediators of ischemia other than NO. Basic Fibroblast Growth Factor
(bFGF) protects against ischemia in both rat and cat models. EA applied
10 minutes after occlusion of the MCA in rats upregulated the
_expression
of bFGF-like immunoreactivity in the striatum and cortex of rat
brains.(8,9) Following transient global ischemia, acupuncture
suppressed
apoptosis in the hippocampus of gerbils.(10) In this study, the
acupoint
Zusanli had the most potent effect.

More limited data suggests a role for acupuncture in treating
hemorrhagic injury. Zusanli applied to rats with intrastriatal
hemorrhage decreased lesion size and apoptotic neuronal cell death.(8)
Studies of EA at acupoint Neiguan suggest that acupuncture exerts a
pressor effect during hemorrhagic hypotension in dogs.(9)

Finally, acupuncture administered to rats at Bai Hui 15 minutes before
a
training trial significantly attenuated cyclohexamide-induced
impairment
of passive avoidance. Efficacy paralleled duration of acupuncture: rats
treated for 60 minutes had the greatest response, and decreasing levels
of response were seen in those treated for 30 and 15 minutes. This
suggests that acupuncture could positively affect the memory storage
system, a system frequently damaged by stroke.

Cumulatively, these studies indicate that acupuncture has the potential
to decrease the volume of infarct and enhance recovery from ischemic
brain injury. The studies that investigated the effects of specific
acupoints provide the rudiments of a scientific rationale for point
prescriptions.

Physiologic studies in Humans
Four studies in humans have attempted to demonstrate physiologic
mechanisms by which acupuncture affects ischemia. In six patients with
MCA occlusion and eight healthy volunteers, single-photon emission
computed tomography (SPECT) brain perfusion images were obtained before
and after acupuncture at LI4, 10, 11, 15, 16 and TE5.
All participants exhibited increased cerebral blood flow (CBF), but in
stroke patients the increase was pronounced in the perilesional and
use-dependent areas.(10) Another study focused on the mechanisms by
which blood flow is stimulated. Endothelin (ET) is a potent
vasoconstrictor which also increases levels of the vasoconstrictor and
platelet-aggregating factor thromboxane B2 (TXB2). In contrast,
6-keto-prostaglandin-F1a (6-keto-PGF1a) is a potent vasodilator and
inhibitor of platelet aggregation. 20 healthy subjects and 20
"convalescent" stroke patients were given body acupuncture 30 minutes a
day for 10 days. The serum level of ET was significantly higher than
that of healthy subjects at baseline. After treatment it declined
significantly but did not normalize. Urine levels of TXB2 and
6-keto-PGF1a were also higher in cases at baseline. TXB2 declined
significantly (but did not normalize), but 6-keto-PGF1a levels did not
change significantly.(11) These results suggest that the net effect of
acupuncture is to increase vasodilation, which could explain the
increased CBF.

Two other physiologic studies in humans attempted to delineate groups
of
patients for whom acupuncture would be effective.  64 healthy subjects
and 16 post-stroke patients were administered scalp acupuncture. After
treatment, those with stroke expressed Motor Evoked Potentials (MEP)
with similar waveforms and intervals to controls, but with lower
voltage, a latent period, and an after effect. In 10 cases there was
damage to the pyramidal tract and basal ganglia. For these patients,
needling scalp on the affected side did not elicit MEP on the
contralateral side, but needling the scalp on the unaffected side
elicited MEP in the ipsilateral (paralytic) hand with increased latent
period, markedly lower amplitude and a shorter interval. This suggests
that even patients with damage to the pyramidal tract may benefit from
acupuncture. Unfortunately, the extent of the damage was not
quantified.(12) Finally, in a study of 64 subjects with acute ischemic
cerebral events, 33 were given routine treatment while 31 were given EA
to body points for 20 minutes a day for 14 days, in addition to routine
treatment. 26 healthy subjects functioned as controls. Both before and
after treatment, the level of plasma somatostatin (SS) in the EA group
did not differ significantly from controls. But among those who
responded well to acupuncture, the levels of SS in serum and
cerebrospinal fluid were significantly higher than among those who did
not respond well. Similarly, the serum level of pancreatic polypeptide
did not differ significantly between groups, but within the acupuncture
group those who responded well had significantly higher levels than
those who did not. 

This type of research, which is not relevant in a review of efficacy,
is
essential to practitioners. Few if any practitioners believe that
acupuncture will be beneficial for everyone who has ever had a stroke.
Practitioners implicitly or explicitly apply selection criteria in
deciding whom to treat. One of the goals of research is to provide
clinicians with selection criteria that guide the choice of treatment
modalities to maximize benefit to patients. Chemical markers of
efficacy
such as the ones explored above, if confirmed, could be among the
indicators that an individual patient would benefit from acupuncture.

Other indicators are suggested in the study reported by Naeser et
al.(13) Her group performed acupuncture on 10 acute and 10 chronic
stroke subjects. They correctly predicted response to acupuncture in 19
of the 20 patients based on CT scan lesion site. Those who responded
well had damage to less than half of the motor pathway areas on CT
scan,
especially in the periventricular white matter area at the level of the
body of the lateral ventricle. 8 of the 20 patients had significant
improvement in motor function, including 3 of the 10 subjects treated
more than 3 months post stroke and 5 of the 10 treated less than 3
months post stroke. Most improvements were sustained for at least 4
months after the last acupuncture intervention. Unfortunately, clinical
researchers have not integrated this type of information into their
study designs.

Clinical Outcome Studies
In the 1990s there was an explosion of promising research on the
clinical outcomes of acupuncture and transcutaneous electric nerve
stimulation (TENS) for stroke.

Tekeodlu and colleagues compared the Barthel Index of Activities of
Daily Living (ADLs) in 30 post-stroke subjects who received
high-frequency TENS and 30 who received placebo TENS. (14) At entry the
intervention group was more disabled (Barthel score 30.4 +/- 22.1
versus
44.7 +/- 17) than the control group. At the end of the study the
intervention group had globally improved significantly more than the
control group. Subjects given the intervention improved in all 10
categories measured on the Barthel Index, whereas in the control group
they had improved significantly only in 5 of the 10 categories.
Specifically, the control group showed no significant improvement in
grooming, feeding, mobility, climbing stairs or bathing.

Sonde and colleagues compared subjects who were 6 - 12 months
post-stroke and received low-frequency TENS to subjects 6 - 12 months
post-stroke who did not receive this additional intervention.(15) They
focused on the functional motor capacity of the paretic arm. The 26
patients in the intervention group had significantly improved motor
performance on the Fugl-Meyer scale, compared to the 18 controls. There
was no improvement in ADLs. Although there was no decrease in pain or
spasticity, this was expected, as high-frequency TENS (which was not
used in this intervention) is the standard type used for pain and
spasticity. Given the lack of improvement in ADLs, it is difficult to
assess the clinical significance of the improvement that was made.
Perhaps to answer this question, a follow-up study was performed 3
years
later. Both groups had declined below baseline on the Fugl-Meyer Motor
Performance Scale and spasticity had increased in both groups, though
not significantly. On the other hand, the ADL score deteriorated
significantly in the control group, but did not change significantly in
the intervention group.(16)

A similar study, with even more favorable results, was conducted by
Wong
and colleagues. 128 subjects within 2 weeks of stroke onset were
randomized to receive comprehensive rehabilitation with or without
electrical stimulation of acupuncture points through adhesive surface
electrodes, 5 times a week for 2 weeks. Neurological status was
assessed
by Brunnstrom's stages and the Chinese version of the Functional
Independence Measure (FIM), before treatment and at discharge. The
group
that received electrical stimulation had a significantly shorter
duration of hospital stay (29.1 +/- 7.9 days vs. 32.4 +/- 8.2 days),
and
scored significantly better on FIM for self-care and locomotion.(17)

A Kjendahl et al compared the responses of subjects given six weeks of
acupuncture treatment during the subacute phase of stroke, to subjects
given the standard rehabilitation program in a rehabilitation unit.
Those who received acupuncture showed significantly more improvement in
motor function and ADLs on discharge, compared with those who did not
receive acupuncture.(18) In a one-year follow-up study (1997) they
found
that although both groups continued to improve, those in the
acupuncture
group improved significantly more on the Motor Assessment Scale,
Sunnaas
Index of ADLs and Nottingham Health Profile. Of note, the points
selected were individualized to each patient, in accordance with
Traditional Chinese Medical Theory.

Another, earlier pair of studies showed equally remarkable results. K
Johansson et al conducted a trial in which 78 stroke patients with a
median age of 75 were randomized within 10 days of stroke onset to
receive either standard care or standard care plus additional sensory
stimulation, including EA. The group receiving EA had significantly
better scores on balance, mobility and ADLs.(19)  In a follow-up study
by Manusson et al (1994) more than 2 years after stroke onset (mean 2.7
years, range 2.0-3.8 years) they investigated postural control in the
48
survivors of the original trial. 22 subjects from the EA group and 26
from the control group were compared with 23 age-matched healthy
subjects. Significantly more subjects in the treatment group than in
the
control group maintained stance during perturbation. The values
approached the normal for age-matched healthy controls.(20) To account
for the persistence of effects so long after the intervention, they
speculate that the sensory stimulation provided by EA enhances the
functional plasticity of the brain.

Two studies of the effects of acupuncture on spasticity produced
contradictory results. 25 patients with chronic poststroke leg
plasticity were randomized to placebo needling (n=12) or real treatment
(n=13). At the end of 4 weeks there were no significant differences
between the two groups on the Modified Ashworth Scale (MAS). 35 stroke
patients with elbow spasticity were randomized to EA (n=15),
moxibustion
(n=10) or control (n=10). The EA group received 30 minutes of
stimulation every other day at LI11, LI10, TB5 and LI4 on the paretic
side. The moxa group received direct moxa three times a day every other
day, to the same points. The control group received "routine
acupuncture" therapy for stroke. All patients also received range of
motion exercises. Significant reductions in spasticity, as measured by
MAS, were achieved in the EA group after day 5, and persisted at day 15
after the start of treatment. There were no significant changes in
either of the other groups.(21)

In contrast to the studies mentioned so far, three well-known clinical
trials produced negative outcomes. Gosman-Hedstrom, et al conducted a
study they state was intended specifically to examine the possible
placebo effects of acupuncture. 104 subjects were randomized to 3
groups: deep, superficial and no acupuncture treatment. Assessments
were
conducted 4 times during the first year after randomization.(22) No
differences were found in changes in neurological score, Barthel or
Sunnaas ADL index scores. On the Nottingham Health Profile the no
acupuncture group had somewhat fewer mobility problems.

The main problem with this study is that all the statistical analyses
were performed in 2 steps. In the first step the superficial and no
acupuncture groups were compared. If no significant differences were
observed, they were combined and compared to the deep acupuncture
group.
This resulted in comparing 37 patients who received deep acupuncture to
67 patients, 34 of whom received superficial acupuncture. Superficial
acupuncture may have a modest, systemic effect, and most acupuncturists
would not consider it an adequate control arm. Combining its effects
with that of the no acupuncture group may have masked a modest
improvement in those who received some kind of acupuncture, compared to
those who received no acupuncture.

The Swedish Collaboration on Sensory Stimulation After Stroke whose
1993
article had been important in attracting attention to acupuncture for
stroke, subsequently published a negative study. 150 subjects with
moderate or severe functional impairments were randomized 5 to 10 days
after acute stroke to 1 of 3 groups: EA, high-intensity, low-frequency
TENS, or low-intensity (subliminal) high-frequency electrostimulation.
At 3-month and 1-year follow-ups, no clinically important or
statistically significant differences were observed in motor function,
ADLs, walking ability, social activities or life satisfaction.(23)

As Schiflett points out in a letter to the editor, what appear to be
clinically important improvements in the acupuncture group, such as a
greater than 100% increase in walking speed, are ignored by restricting
analysis to increases in global scores in quality of life. Other
improvements are masked by using nonparametric statistics and intention
to treat analysis (which included assigning a functional score to
patients who died as a result of disease, not as a result of
acupuncture).

In 2002, Sze published a prospective randomized controlled trial on 106
patients in a rehabilitation unit enrolled 3 to 5 days after acute
stroke. The control group received standard modalities, including
physiotherapy, occupational and speech therapy. The intervention group
received manual body acupuncture at LI4, 10, 11, and 15, TE5, GB30 and
34, S36 and 41, with optional points CV4, 6, 10 and 12, and S24 and 26.
The treating acupuncturist was allowed to omit or add a maximum of 3
acupoints. Acupuncture was administered on the paretic side for 30
minutes per session 5 times a week for inpatients and 3 times per week
for outpatients, for 10 weeks. A mean of 35 treatments were received.
No
differences in Fugl-Meyer Assessment (FMA), Barthel Index (BI) or
Functional Independence Measure (FIM) were noted at weeks 0, 5 or
10.(24)

The power analysis for this study indicated that 40 patients were
required for each arm to have 0.8 power to detect an effect size of
0.5.
However, the study stratified patients in each arm into two groups -
those with moderate and those with severe levels of disability.
Although
53 patients were treated with acupuncture, only 31 of them had severe
stroke and 22 had moderate stroke. Neither group is large enough to
achieve the requisite power. Grouping them together makes the
assumption
that acupuncture will be just as effective for moderate as for severe
stroke. No justification is given for this assumption. In addition,
despite their statement that the groups were evenly matched in baseline
characteristics, table 1 appears to show significant differences in CT
scan result, number and location of lesions. Finally, in the severe
group, scores on the BI and FMA (total), and change in median FMAM
(motor) are in fact higher in the treatment group. In the treatment
group, the FMAM (motor) median score improved from 29.7 to 51.2, while
in the control group it improved from 38.7 to 53.3. The interquartile
range in the treatment group (14.0 - 46.4) is much narrower than in the
control group (14.1 - 64.5). It is possible that in this underpowered
study, outliers in the control group skewed the results, making it
appear that improvement in the two groups was equivalent.

Later the same year Sze and colleagues published a meta-analysis of the
effect of acupuncture on motor recovery after stroke. The review
concluded that acupuncture had no additional effect on recovery in
patients who were given stroke rehabilitation, and that it had a
positive effect on motor recovery in patients who did not receive
rehabilitation and were treated at either an unknown interval or more
than six months after stroke onset. These conclusions were reached by
pooling data from heterogeneous trials, many of which are summarized
above. Pooling required a large number of statistical conversions,
which
may have obliterated clinically important aspects of the trials. As
discussed above, restricting analysis to global measures of improvement
may mask clinically significant improvements in specific components of
the global measures.(25)

Lessons and Directions for Future Research
Perhaps the main lesson we can learn from the experience of trying to
prove efficacy of acupuncture for stroke is the necessity for
preliminary studies that attempt to define the set of patients for whom
acupuncture is most likely to be beneficial. Since CT scans are
routinely performed in stroke patients, a reasonable next step would be
to compare and contrast retrospectively the CT scans of those who did
and did not improve after acupuncture. An efficacy trial can be
designed
subsequently. Such a trial may well show an additional benefit to
acupuncture in addition to rehabilitation. The second aspect of
analysis
missing from most acupuncture research is cost effectiveness. If Sze
and
colleagues turn out to be correct that acupuncture adds little to
comprehensive rehabilitation programs but is helpful in their absence,
then in areas without the resources to establish such programs
acupuncture may be a viable alternative. Just as importantly, if
acupuncture is as effective as rehabilitation programs, then it may be
a
viable, lower cost alternative. Assuming that efficacy can be
demonstrated for some patients, cost effectiveness analysis, which is
largely absent from acupuncture research, will be an essential piece of
the argument that its widespread use should be adopted.

References:

1. Rabinstein AA, Sluman LM. Acupuncture in clinical neurology
[Review][120 refs] Neurologist. 2003; 9(3):137-48

2. Ying SX, Cheng JS, Jin ZQ, Cheng JS. Acupuncture for Stroke:
Physiologic Mechanisms Studied in Animals 1994, 1997

3. Zhao P. Acupuncture for Stroke: Physiologic Mechanisms Studied in
Animals 2000

4. Yang R, Huang ZN, Cheng JS. Acupuncture for Stroke: Physiologic
Mechanisms Studied in Animals 2000

5. Wei GW, et al. Acupuncture for Stroke: Physiologic Mechanisms
Studied
in Animals 2000

6. OuYang W, et al. Acupuncture for Stroke: Physiologic Mechanisms
Studied in Animals 1999

7. Jan MH, Shin MC, Lee TH, et. al. Acupuncture suppresses ischemia
induced  increase in c-Fos _expression and apoptosis in

the hippocampal CA1 region in gerbils. Neuroscience Letters. 2003 Aug.
14; 347(1):5-8

8.  Cho NH, Lee JD, Cheong BS, Choi DY, et al. Acupuncture suppresses
intrastriatal hemorrhage induced apoptotic neuronal  cell death in
rats.
Neuroscience Letters. 2004 May 20; 362(2):141-5

9. Syuu Y., Matsubara H., Hosogi S., Suga H. Pressor effect of
electroacupuncture on hemorrhagic hypotension. American  Journal of
Physiology - Regulatory Integrative & Comparative Physiology.2003 Dec.;
285(6):R1446-52 (abstr)

10. Lee JD, Chon JS, Jeong HK, Kim HJ, Yun M. Kim DY, Kim DI, Park CI,
Yoo HS. The cerebrovascular response to traditional acupuncture after
stroke.  Neuroradiology. 2003 Nov.; 45(11):780-4

11. Zhang S. Acupuncture for Stroke: I. Pathophysiologic Studies:
Animal
and Humans 1999

12. Sun and Sun. Acupuncture for Stroke: I. Pathophysiologic Studies:
Animals and Humans 1998,

13. Nasser M, Alesander, MP, Stiassny-Eder D, et.al. Acupuncture and
Electro-Therapeutics Research 1994 Oct.-Dec.; 19(4):227-49.

14. Tekeodlu Y. Acupuncture for Stroke: Clinical Outcome Studies in
Humans 1998

15. Sonde L, et al. Acupuncture for Stroke: Clinical Outcome Trials in
Humans 1998

16. Sonde L, et al. Acupuncture for Stroke: Clinical Outcome Trials in
Humans 2000

17. Wong AMK, et al. Acupuncture for Stroke: Clinical Outcome Trials in
Humans 1999

18. Kjendahl A, et al. Acupuncture for Stroke: Clinical Outcome Trials
in Humans 1996

19. Johansson K, et al. Acupuncture for Stroke: Clinical Outcome
Studies
in  Humans 1993

20. Magnusson, et al. Acupuncture for Stroke: Clinical Outcome Studies
in  Humans 1994
        
21. Fink M., Rollnik JD, Bijak M, Borstadt C, Dauper J., Guergueltcheva
V., Dengler R., Karst M. Needle acupuncture in chronic poststroke leg
spasticity. Archives of Physical Medicine & Rehabilitation 2004 Apr.;
85(4):667-72
        
22. Gosman-Hedstrom, et al. Acupuncture for Stroke: Clinical Outcome
Studies in Humans 1998

23. Johansson BB, Haker E, von Arbin M, et.al. Acupuncture and
Transcutaneous Nerve Stimulation in Stroke Rehabilitation: A
Randomized,
Controlled Trial.  Stroke 2001 March; 32(3):707-713

24. Sze SK, FRCP, Wong E, et.al. Does Acupuncture Improve Motor
Recovery
After Stroke? A Meta-Analysis of Randomized  Controlled Trials.
Acupuncture  for Postroke Motor Recovery 2002; 33:2604-2619

25. Ibid